Advanced Multi-Modal Sensor Fusion System for Detecting Falling Humans (AFODS) - Source Code Archive v1.0.1

Collisions with fallen pedestrians pose a lethal challenge to current Advanced Driver Assistance Systems (ADAS). Standard visible-spectrum systems yield a True Positive Rate (TPR) as low as 21.4% at night (0 lux) for prone individuals — a critical and life-threatening classification gap that no current regulatory test protocol addresses. This repository contains the software framework and quantitative evaluation for the Advanced Falling Object Detection System (AFODS), supporting the peer-reviewed manuscript: Barua, N. & Hitosugi, M. (2025). Advanced Multi-Modal Sensor Fusion System for Detecting Falling Humans: Quantitative Evaluation for Enhanced Vehicle Safety. Vehicles, 7(4), 149. https://doi.org/10.3390/vehicles7040149 AFODS architecturally integrates LWIR Thermal, NIR Stereo, and Ultrasonic sensors, processed through a custom AI pipeline combining YOLOv7-Tiny for object detection and a GRU-based RNN for proactive threat assessment. Validated across 320 controlled trials, the system achieved: - Clear Daylight TPR: 99.5% (Baseline: 96.8%, p = 0.041)- Night (0 lux) TPR: 98.2% (Baseline: 21.4%, p < 0.001)- Rain (50 mm/h) TPR: 96.4% (Baseline: 55.7%, p < 0.001)- Fog (